Study on Data Compression for TDOA Estimation in WSN Application

Data compression is the key method to slove the conflicts between the limited sensor energy and communication capability in wireless sensor networks based passive location system. In this paper, a survey on the researches about compression for estimation is given, and the Fisher-information-based data compression method for TDOA estimation application is particular introduced. In order to improve the Lagrangian Multiplier based bit allocation method which the original data compression method uesd is not efficient, a rapid bit allocation method based on the proportion of Fisher-information each frequency components took in the total Fisher-information is proposed; In addition, due to the amplitude of the signal have no relation with the time delay, sending only the phase of the signal instead of the real and imaginary parts of the DFT coefficients is proposed. The simulation results show that the new bit allocation method’s process time compare to the Lagrangian Multiplier based bit allocation method and the water-filling based bit allocation method is reduced, and only leads slight TDOA estimation accuracy’s deterioration; In the same compression ratio, sending only the phase of the signal acquire a better TDOA estimation accuracy in high SNRs than sending the real and imaginary parts of the signal’s DFT coefficients. Overall, the improved method is more operational than the original one.